JABS Documentation

JABS (JAX Animal Behavior System) is a machine learning-based tool for automatically identifying and classifying animal behaviors in video recordings. It uses pose estimation data to extract features and train classifiers that can recognize specific behaviors across multiple animals and videos.

Key Features:

• Train custom behavior classifiers with manual labeling
• Support for single and multi-animal pose estimation
• View prediction visualization during labeling to support active learning
• Batch processing via command-line tools
• Export classifiers for batch processing